Methods of Modeling Transport Demand

Real time traffic load forecasting

“…Computational intelligence, engaging hybrid SSA-ANN was an effective approach for modeling complex forecasting phenomena (Profillidis and Botzoris, 2018;Andersson et al, 2017). Hybrid methodology provided a robust potential for modeling, analyzing and forecasting, compared to conventional ANN models (Vlahogianni et al, 2005;Profillidis and Botzoris, 2018).…”

“…• Hybrid SSA -ANN methodology could improve transportation system management. (Profillidis and Botzoris, 2018;Andersson et al, 2017). Οι υβριδικές προσεγγίσεις προβλημάτων πρόβλεψης μπορούν να βελτιώσουν την ήδη πολύ καλή προβλεπτική ικανότητα των Τεχνητών Νευρωνικών Δικτύων (Vlahogianni et al, 2008;Profillidis and Botzoris, 2018…”

“…Similarly, a univariate time series referred to a time series that consisted of only one variable, whose data or observations were recorded sequentially, usually over equal time intervals. Although univariate time series were usually considered as a series of observations, time was in fact an implicit variable that was a variable not explicitly, not plainly expressed (Profillidis and Botzoris, 2018). The need for univariate time series analysis was appropriate for situations where:…”

“…3 presented monthly traffic data for motor bikes at toll station located at Moschohorion, near Larissa, Greece. Another category of stationarity tests was the unit root tests, the most popular among these being the Augmented Dickey-Fuller test (Profillidis and Botzoris, 2018). Σχήμα 3: Χρονοσειρά που περιγράφει μηνιαίο μεταφορικό φόρτο.…”

“…The only independent variable in time series methods was time. All other variables affecting transportation characteristics were considered in time series methods to pursue the same path in the future as in the past, and to continue over the course of time to affect transport demand in the same manner and with the same degree of influence (Profillidis and Botzoris, 2018).…”